Relationship compatibility test: what research says matters

A relationship compatibility test can show how your relationship looks to you right now, in areas like communication, conflict and affection. It can’t forecast how the relationship will change: in a 2020 analysis of 43 couples studies, changes in relationship quality were largely unpredictable from people’s questionnaire answers.

That gap matters at the moments when people go looking for a score: six months after signing a lease, in the middle of a running argument over whose turn it is to do the dishes, or on the drive home from a weekend with the in-laws. Many of the studies involved married or dating couples made up of a man and a woman, and the samples are described where they matter.

What can a relationship compatibility test tell you?

A relationship compatibility test can describe how the relationship feels to you at this point, but a study its authors called the largest data-analytic effort in the study of romantic relationships found that questionnaire answers did a poor job of forecasting how relationships would change.

That study, published in the Proceedings of the National Academy of Sciences in 2020 by Samantha Joel and dozens of co-authors, drew on 11,196 couples from 43 longitudinal studies run by 29 research labs. The team used machine learning to test how well the answers people gave at the start of each study predicted how satisfied and committed they were, both then and later. The answers predicted where satisfaction stood at the start of a study better than where it stood at the end, and they did poorly at predicting how it would rise or fall: no analysis of that change accounted for more than 5% of the variance, typically over one to two years. Change in commitment was, in the paper’s words, “generally unpredictable.”

The authors offered a guess about why. Change, they wrote, “is likely a function of external context, behavioral processes, or other factors that are themselves changing over time.” A layoff, a cross-country move for a new job or a parent who gets sick are the kinds of outside events the authors’ explanation points to. The quiz page makes a similar point in plainer words: “Your result reflects how you answered this time.”

Compatibility by zodiac sign, name or birthday is a separate question, and our life path compatibility guide looks at the birthday-number version of it.

What actually predicts a happy relationship?

In the same 2020 analysis, how each person saw the relationship accounted for up to 45% of the variation in their satisfaction at the start, and adding what the partner reported, or either person’s personality and traits, did not improve that prediction.

The specific judgments that topped the list, such as appreciation and how committed the partner seemed, are covered in our life path compatibility guide.

“How each person saw the relationship” means their own ratings of things like conflict and affection. The partner’s reports were what the partner said in the same studies, on their own questionnaires, about themselves and the relationship. A person’s own view of the relationship predicted 45% of their satisfaction at the start of a study, and what the partner reported about the relationship predicted 15%. With the partner’s reports added, the figure came to 46%, and with every measure in one model it was 44%, which the authors described as adding “no predictive power” beyond a person’s own view. By the end of the studies, a person’s own view still predicted up to 18%.

Personality and other individual traits counted for something on their own. Measures such as neuroticism, age and gender predicted up to 21% of relationship quality at the start and as much as 12% at the end. They added little or nothing once a person’s view of the relationship was already in the model. The authors did not take this to mean a partner’s feelings or traits are unimportant. One possibility they raised is that those factors work indirectly, by shaping the day-to-day experience of the relationship that a person’s own ratings already capture.

The paper notes that the edge for relationship measures over individual traits fits earlier reviews of the field. One of those, by Benjamin Karney and Thomas Bradbury in 1995, reviewed 115 longitudinal studies representing more than 45,000 marriages.

The authors also spelled out the study’s limits. All 43 datasets came from Western countries: the United States, Canada, Switzerland, New Zealand, the Netherlands and Israel. Of the 1,149 relationship measures tested, 99.4% were self-report rating scales rather than observations of how couples actually behaved. Outside factors such as income and stress were self-reported too, so the results don’t speak to records like tax returns or biological stress measures. More of the datasets came from psychologists’ labs than from sociology, communication or family studies. And the change being predicted was usually tracked over one to two years.

Do you need to be alike to be compatible?

Being alike matters less than it sounds: in studies of existing relationships, feeling similar was linked to attraction, while measured similarity was not.

A 2008 meta-analysis by Montoya, Horton and Kirchner pooled 460 effect sizes from 313 studies. Across all of them, both actual similarity, meaning how alike two people really were, and perceived similarity, meaning how alike they believed they were, went with attraction. Actual similarity did matter in studies where people had little or no interaction. The split showed up among people who already knew each other well: in existing relationships, actual similarity no longer predicted attraction, while perceived similarity still did.

A separate 2010 study by Dyrenforth and colleagues, using nationally representative samples from Australia, the United Kingdom and Germany, found that once the effects of each partner’s own personality were accounted for, personality similarity explained less than 0.5% of the variation in relationship satisfaction.

In everyday terms, one partner plans the grocery list by aisle and the other shops by whatever looks good. On paper they look like opposites, and they can feel like they’re on the same page about what matters. The similarity meta-analysis measured attraction rather than long-term happiness, so it says more about feeling drawn to someone than about how a relationship will go.

Do matching algorithms work?

As of 2012, psychological scientists who systematically reviewed the evidence concluded that dating sites’ matching-algorithm claims were unsubstantiated. A 2017 speed-dating study found that traits and preferences measured before people met couldn’t predict which two people would be drawn to each other.

The 2012 conclusion comes from a news release by the Association for Psychological Science about an article in its journal Psychological Science in the Public Interest. “To date, there is no compelling evidence that any online dating matching algorithm actually works,” Eli Finkel, the article’s lead author, said in the release. The release also pointed out that the strongest predictors of relationship well-being, “such as a couple’s interaction style and ability to navigate stressful circumstances, cannot be assessed with such data,” meaning information about two people gathered before they meet.

The 2017 study, by Samantha Joel, Paul Eastwick and Eli Finkel, put that idea to a direct test across two speed-dating studies. Participants filled out more than 100 measures of their traits and preferences ahead of time. Machine learning could use those answers to predict how much a person tended to desire others and be desired in general, but not the attraction unique to a particular pair.

A matching algorithm is guessing before two people have ever had a conversation. Someone looking back at a relationship they are already in is answering a different kind of question.

Arguing isn’t the problem; how you repair is

Couples who argue aren’t doomed: research links hostility and a lack of warmth to worse outcomes.

A 2011 meta-analysis by Erica Woodin, covering 64 studies and 5,071 couples, found that hostility during conflict was consistently linked to lower relationship satisfaction, with a medium-sized effect. Intimacy and problem solving during conflict were closely tied to higher satisfaction, also at a medium size, and men and women differed only slightly in how often they showed these behaviors.

Timing shows up in a 14-year study by John Gottman and Robert Levenson, published in 2000. Negative emotion during conflict predicted divorces that came early in a marriage. A lack of positive emotion, in everyday conversation as well as in conflict, predicted divorces that came later.

A 2015 paper by John Gottman, Janice Driver and Amber Tabares, on newlywed couples, discusses repair attempts once a conflict has started. Practical ways to make up after a fight are in our guide to couples arguing.

Some arguments are not about communication at all. When fights bring fear, control or harm, the problem is no longer how two people talk, and our guide covers that in its section on when arguing stops being a communication problem.

The five areas the quiz looks at, and what research says about each

The quiz gives a 0–100 score across five areas: communication, conflict and repair, shared direction, affection, and chores and money. Each is an area relationship researchers have studied.

One person takes it, answering 30 questions about the relationship as they see it, and the result comes with a relationship radar, three conversation starters and one weekly ritual. Each row pairs one area with one or two studies on the same topic.

This article doesn’t describe how the quiz scores each area, and the studies in the table are research on the same topics rather than the basis of the quiz.

Communication. In a 2010 study led by Howard Markman, couples planning to marry were recruited through the religious organizations performing their weddings, as part of a study of a premarital education program. Negative communication before the wedding, whether observed by the researchers or reported by the couples themselves, was associated with lower marital adjustment over the first five years. Positive communication observed before the wedding was not. For divorce, self-reported negativity was significantly related, while the observed measures only approached significance, and the authors noted that the small number of divorces limited what the study could detect.

Shared direction. In a study of 105 dating couples, Judith Gere and Ulrich Schimmack found that the more people reported their goals conflicting with their partner’s, the lower their relationship quality and well-being. It is a correlational study, so it shows that the two were linked, not that one caused the other. One example is a partner angling for a job offer in Denver while the other has just put down roots near family in Ohio.

Affection. A 2013 study by Anik Debrot and colleagues had 102 dating couples report in an electronic diary four times a day for a week. Touch in response to a partner’s mood, such as a hug or other affectionate touch, was associated with a better mood in the person being touched, and that link ran through a sense of psychological intimacy. People who were touched more during that week also reported better psychological well-being six months later. How people give and receive care in other ways is the subject of our love languages guide.

These studies describe patterns across groups of couples, and none of them can place a particular relationship on a scale. Your score reflects the relationship as you see it from your side.

Chores and money: fairness, and the fights that repeat

With chores, whether the split feels fair seems to matter; with money, the arguments aren’t the most common, but they’re the ones that keep coming back.

Daniel Carlson and colleagues looked at two surveys of couples, one from 1992–1994 and one from 2006. How couples divided housework was tied to their relationship differently in the two periods, and the authors attributed the shift to perceived fairness playing a bigger role over time. Attitudes moved the same way. In a 2007 Pew Research Center survey, 62% of American adults said sharing household chores was very important to a successful marriage, up from 47% in 1990, putting it third among nine items. That was a survey of opinions in 2007, not evidence about which marriages last.

Money looks different. In a diary study by Lauren Papp, Mark Cummings and Marcie Goeke-Morey, 100 married couples with at least one child aged 8 to 16, mostly white, recorded their disagreements at home in 1999–2000. Children were the most common topic, followed by chores, communication and leisure. Money ranked sixth in husbands’ diaries and fifth in wives’, but money conflicts were more likely to be a recurring problem and more often stayed unresolved. For many couples now, that might be a rent split through a payment app that never quite feels even. More on couples and money is in our money personality guide.

How to use your result: a conversation, not a verdict

A compatibility score is a starting point for a conversation, not a verdict on the relationship.

The result page includes three conversation starters and one weekly ritual. The studies here touch on the same ground: how the last argument got repaired, whether the chore split still feels fair, or where two sets of plans are pulling in different directions. Because changes in relationship quality were so hard to predict in the 2020 analysis, a score describes a moment rather than a destiny.

For a partner who wants in, there are the Love Language Quiz and the Communication & Conflict Style Test: each person takes each one once, and the talk afterward is about what each of them got.

If trust and closeness keep coming up, our guide to attachment styles looks at where those patterns come from.

A score can start a conversation that might otherwise wait for the next fight. Whatever you do with it, your result shows the relationship as you answered it today.